[time-nuts] Thinking outside the box a super reference
Richard (Rick) Karlquist
richard at karlquist.com
Sat Nov 5 11:32:58 EDT 2016
One of the main limiting factors in the 5061 was
microwave leakage. An excellent Italian engineer
named DiMarchi mastered the so called "top cover
effect", where removing the top cover changed the
frequency. He had a small business going refurbishing
5061's by cleaning up the waveguide gasketing, etc.
If any 9192.63177 reaches the beam at one end or
the other, it will upset the phase balance. In
the 5071, phase balance is the main limiting factor
in accuracy. They go to extreme measures to make
the cavity absolutely symmetrical using fabrication
techniques analogous to "self aligning" IC masking.
In the 5071, the only place 9192 shows up is in the
microwave module that is directly attached to the
coax to waveguide transition into the cavity.
There are no frequencies anywhere that are sub
harmonics of 9192. Incidentally, there are no
frequencies anywhere that are coherent with
50 Hz, 60 Hz, etc line frequencies. Nothing
is by accident when Len Cutler is involved.
In terms of basic synthesizer architecture, the
mix from 9280 to 9192 using 87 is described by
the technical term "free lunch" :-) We pick
up two decades of resolution. Furthermore, we
don't have to filter out 9280 or 9367 because
they are ignored by the CBT. One of the reasons
for going up from 12 to 87 was to get these
spurs safely removed from anything that would
interact quantum mechanically with the cesium
line tail. With the increased accuracy, 12
was no longer high enough.
On 11/5/2016 7:39 AM, Attila Kinali wrote:
> Hoi Rick,
> On Sat, 5 Nov 2016 07:17:21 -0700
> "Richard (Rick) Karlquist" <richard at karlquist.com> wrote:
>> I think this is all described in the 1992 FCS papers,
>> but the executive summary is that a direct synthesizer
>> on 9192.63177 is to be avoided at all costs because
>> of the danger of it leaking into the CBT cavity.
>> This is also the reason why you don't multiply up
>> a subharmonic of this frequency.
> I don't get what you mean with "danger of leaking into the CBT cavity"?
> When signal leakage into the cavity is a problem, shouldn't that also
> exist for the signal after the mixer? And what does this leaking actually
> mean? The 9192.63177 is supposed to end up in the cavity anyways.
> Attila Kinali
More information about the time-nuts